Hundreds of susceptibility genes have been identified for autism spectrum disorder (ASD), and many are related to synaptic function. This has led to a hypothesis that the deficits in ASD may reflect an imbalance in the relative contributions of excitatory and inhibitory synaptic inputs. Canonical neural computations are stereotyped, modular circuit functions that occur across the brain and can provide building blocks for more complex operations. Disruptions to these computations would be expected to have negative behavioral consequences. Interestingly, divisive normalization, one such canonical neural computation, computes a ratio between individual neuronal responses and the summed population activity, and inherently reflects the balance of excitation to inhibition.

Dora Angelaki and her colleagues seek to test the hypothesis that canonical neural computations are altered in ASD. Previously, Angelaki’s laboratory showed that an imbalance in excitation and inhibition, due to a reduction in the amount of inhibition occurring through divisive normalization, could account for the perceptual alterations observed in individuals with ASD¹. Now, Angelaki’s team proposes to use a divisive normalization model of primary visual cortex to test previously published psychophysics results and explore whether tasks that are believed to require divisive normalization are altered in ASD.

The first set of experiments will include two cohorts of individuals with defined genetic mutations (Rett syndrome and fragile X syndrome) and another will consist of a large number of high-functioning individuals with ASD.  Individuals will be assessed as they perform simple orientation tasks that are thought to rely heavily on early visual processing and are likely to engage divisive normalization. The genetically defined cohorts will allow Angelaki’s team to characterize function at the level of a single disorder, ideally each with a single set of parameters. The latter cohort will allow the team to assign parameters to individual subjects and test whether these parameters remain consistent across tasks. Angelaki also proposes to extend, in the high-functioning ASD group, both the model and experiments to visual-auditory multisensory experiments that also require divisive normalization.

The outcome of these experiments will aid in our understanding of whether altered canonical computation, possibly of diverse biophysical or genetic implementation, represents a hallmark of ASD symptomatology. This comprehensive, quantitative approach is important for understanding behavioral alterations in ASD and their links to molecular biomarkers. Through this work, Angelaki and her team aim to provide a unifying framework to consolidate behavioral data, as well as take a major step toward establishing a quantitative, testable circuit mechanism linking molecular and behavioral manifestations of ASD.